Since entropy is a state function, ΔS depends only on the initial and final states, not on the route taken between them. Therefore if you can find a reversible path, and calculate the heat flow and thence the entropy change, the entropy change of the
system will be the same even if you take an irreversible path between the same two states.
Let's consider your scenario: two heat reservoirs which we'll call A and B, at temperatures of 100K and 200K respectively.
Suppose that, keeping A and B separate, we bring into contact with them two more heat reservoirs, C into contact with A and D with B, with temperatures 100+δ and 200-δ respectively, where δ is very small, so that heat flow is effectively reversible, and we allow a quantity of heat q to flow from C to A, and from B to D. Then we take away reservoirs C and D. The net result is that we have transferred q reversibly from B to A , with a reversible entropy change ΔS
system = q/100 - q/200 = q/200. As regards the surroundings, we have transferred q from C to D, with a change of entropy q/(200-δ) - q/(100+δ) ≈ -q/200*(1-5δ/200). We can make δ as small as we like, so as δ
0, ΔS
surr -q/200 and ΔS
uni 0.
Now consider your original procedure transferring the heat from B to A directly. The initial and final states of the system (A and B) are the same as in the scenario above - heat q has been transferred from B to A. Therefore the entropy change of the system is the same, q/200. But this time we have done nothing to the surroundings, so ΔS
surr = 0 and ΔS
uni = q/200. So (to answer your question) this is an irreversible process, but the entropy change of the
system (not the surroundings) is the same as for a reversible process.